Re: next-error

[Kevin Rodgers]

Juri Linkov wrote:
> I have some changes not yet installed in CVS (I will install them in
> the next few days) and one of them is changing the rules for finding
> the appropriate source buffer for `next-error'.  Perhaps, this change
> will fix your problem.  Could you provide more details about this case?

Sure.  Here's what the user reports while debugging my
igrep-next-error function (which contains a call to next-error):

        This just doesn't make sense to me. When I call next-error and
        then do a M-: (current-buffer) the correct source buffer is
        displayed as the current-buffer. However, when I capture the
        "current buffer" value after calling next-error from within
        igrep-next-error, it shows to be the *igrep* buffer.

> From what buffer do you call `next-error'?

The problem occurs when the user calls next-error from any buffer
except the *igrep* compilation buffer.

> What was the command that produced the last next-error capable
> buffer?

Only one such command was executed, which calls compile-start in
exactly the same way as M-x grep does.

> Maybe, you need to change the rules in the `next-error-find-buffer'
> function?

Why?  IIUC that function is used to get the compilation buffer, not
the source buffer.  There doesn't seem to be a problem finding that
buffer, as it is highlighted by M-x next-error according to
compilation-start's HIGHLIGHT-REGEXP arg.

Thanks,
--
Kevin Rodgers